Pekka Immonen, Manager, Plant Optimization, OPTIMAX Advanced Control and Optimization of Power Plants
OPTIMAX Suite for Real-Time Optimization Fleet-wide Optimization Unit and Component Optimization Performance Monitoring Combustion Instrumentation Slide 2
FLEET-WIDE OPTIMIZATION Unit Commitment Economic Dispatch Emissions Control Hydro Scheduling Slide 3
FLEET-WIDE OPTIMIZATION Automatic Generation Control Load Sharing Coordinated Frequency Control Slide 4
UNIT AND COMPONENT OPTIMIZATION Using Advanced Process Control (APC) to Improve Stability Reduce Variability Optimize w h y SP Control Algorithm + + + + Σ PLANT Σ y u x Estimator Slide 5
UNIT AND COMPONENT OPTIMIZATION Boiler Start-Up Optimization Unit Response Optimization Coordinated Boiler-Turbine Control Model-Based Runbacks Combustion Optimization Advanced Temperature Control Sootblowing Advisor Slide 6 Faster Start-Up Times Faster Ramp Rates Reduced Emissions Improved Heat Rate Increased Capacity
FLEET-WIDE OPTIMIZATION Optimized Scheduling of Power Generation Short-Term Optimization, time span from hours to days Uses mathematical techniques such as Linear Proramming (LP) and Mixed Integer Programming (MIP) Optimize procurement and generation based on demand forecast Minimize startup and operating costs Optimization of hydro reservoirs Compare different scenarios, and adjust the generation schedules accordingly Provides decision support for users or automatically dispatches generation targets to unit control Slide 7
FLEET-WIDE OPTIMIZATION Energy management for PVO Power Utility Operating a 3.4GW power portfolio including 21 thermal power plants 2 nuclear units 12 hydro stations 5 wind power stations Approximately 25% of total electric production in Finland ABB Process Industries - 8 Slide 8 Pohjolan Voima (PVO) is a privately owned group of companies in the energy sector, which produces electricity and heat at cost for its shareholders in Finland. The Group also develops and maintains technology and services in its sector. PVO-Pool Oy, a subsidiary of Pohjolan Voima Oy, is responsible for the optimal power procurement and the related electricity trade for Pohjolan Voima Group. (www.pvo.fi)
UNIT AND COMPONENT OPTIMIZATION How APC Improves Performance Reduce Improved Control Reduces Process Variance Variability Reduced Variability Allows Shift Operation Closer to Optimum Target Constraints before APC actual constraint Operating closer to limits reduces operating costs, increases production, and/or reduces emissions & wastes, all of which improve profits. Slide 9
uses MPC = Model-Predictive Control: Due to the many constraints and complex interactions, power plant APC and optimization is well suited for the multi-variable model-predictive controller (MPC) The knowledge of dynamic process behavior is contained in the models of the MPC - i.e. it knows the dynamic behavior and all the interactions between the process quantities. The MPC controller evaluates model output response to model input changes, predicts immediate future process behavior, and controls all the variables at the same time, in a coordinated and anticipatory fashion. Examples of COS models: NOX prediction from Overfirea Air (OFA) moves NOX prediction from mill bias moves NOX prediction from burner tilt moves This is demonstrably superior to traditional control (e.g. PID control), where each controller has one input and one output and process dynamics are unknown by the controller. Slide 10
Step Increase in Boiler Coal Input Dynamic Models and Interactions of the MPC Model Inputs Boiler Consumed Air Increases Boiler Steam Flow Increases Model Outputs Temperature Increases Steam Pressure Increases Slide 11
Implementing MPC Solutions NO X Predicted from Overfire Air Moves Model Identification example- NO X vs. Overfire Air Actual NO X Predicted NO X Slide 12
Implementing MPC Solutions Model Identification example- NO X vs. Mill Biases Actual NO X Predicted NO X NO X Predicted from Mill Bias Moves Slide 13
Implementing MPC Solutions Model Identification example- NO X vs. Burner Tilts Actual NO X Predicted NO X NO X Predicted from Burner Tilt Moves Slide 14
MPC Coordinates the Control of Multiple Variables Simultaneously Typical Power Plant Constraints: Maximum Live Steam Temperature Maximum Live Steam Pressure Maximum Reheat Temperature Minimum Flue Gas O 2 Content Maximum Opacity NO X Constraint Condenser Vacuum FD Fan Capacity ID Fan Capacity Windbox Pressure Differential Metal Temperature Gradients In a complex optimization solution, it is necessary to handle multiple controls & constraints simultaneously! Slide 15
Boiler startup optimization Features: Non-linear, Model Predictive Control (NMPC) The boiler model can be identified using measured data during startup phase Parallel calculation of setpoints for fuel, superheater and reheater steam temperature and pressure Explicit consideration of the thermal stress of thick-walled components (controlling temperature differences) Benefits: Equipment protection Shorter startup time reduces fuel consumption Faster load response to load dispatcher, better position for energy trading Possibility for earlier warm-up of steam turbine and grid synchronization Slide 16
Traditional vs. BoilerMax Fuel flow Advanced Control and Optimization of Power Plants Boiler startup optimization achieved benefits HP-Bypass position LS-Temperature dt thickwalled component Savings Traditional startup procedure (dotted lines): Conservative startup to stay within design limits Optimized startup: Exploitation of design limits for more efficient process control Based on process model Predictive coordination of multiple variables Online optimization in real-time Technology: NMPC (Nonlinear Model Predictive Control) Typical 10 to 20% savings possible unused potential design limits Slide 17
Boiler startup optimization achieved benefits Slide 18
Combustion optimization Combustion Optimization is an APC application that automates many combustion control actuators that are normally controlled manually The primary objective of COS is to improve heat rate and lower emissions COS works with base DCS controls COS works with many actuators not tied to feedback in the DCS and also biases DCS PID setpoints Slide 19
Combustion optimization LEGEND MV CV Manipulated Variable Constraint Variable FF Disturbance Feed -Forward MAIN STEAM G LOAD DMD FF RH STEAM RH SPRAY CV WEST TILT DMD MV EAST TILT DMD MV FEEDWATER OFA E - W BIAS MV OFA DMD MV MILL H BIAS MV AUX A -H BIAS MV OPACITY CV MILL G BIAS MV WEST FURNACE EAST FURNACE AUX B -G BIAS MV West AH East AH NOx CV AUX E - W BIAS MV O2 BALANCE CV ID MILL B BIAS MV MILL A BIAS MV WB -FURN DP MV O 2 SP BIAS MV Slide 20 FD
Combustion optimization NOx Reduction (step 1) NOx [ppm] As limit is lowered, more actuators help out: All actuators push back toward their soft targets OFAD and Aux Air reach limits Aux Air bias -5 to 20% OFAD 0 to 100% WindBoxP -1 to 0.5 Slide 21
NOx Combustion optimization NOx Reduction (step 2) [ppm] As limit is lowered, more actuators help out: Tilts and Elevation increase effort after air variables reach limits East Tilt -5 to 15 West Tilt -5 to 15 Mill Elev -10 to 10 Slide 22
Combustion optimization O 2 balancing, Heat Rate Perfect O2 balancing while controlling NOx and minimizing reheat spray O2 A & B OFAD Bias Aux Air Bias Slide 23
MPC VARIABLES: Advanced Control and Optimization of Power Plants Example: Boiler-Turbine optimized control Manipulated Variable Throttle Fuel Spray Demand Valve Flow SP Demand STEAM PRES CV STEAM FLOW CV STEAM TEMP CV Controlled Constraint Variable Generator Steam ID FD ID FD Steam Fan Fan Temperature Pressure Demand Flow MW Current SPRAY DMD MV VALVE DMD MV GEN MW CV PFB FT STOICH AIR PV AT AH FDF FDF DMD CV FDF AMPS CV TG2 G IDF A B C D E FUEL DMD MV Slide 24 IDF DMD CV IDF AMPS CV Stoichiometric Air Flow Process Variable
MPC Tightens Control of MS Temperature & Pressure MPC ON MS Temperature MPC ON MS Pressure Slide 25 Reduce Variability & Increase Targets Improve heat rate and capacity
Energy Efficiency and Performance Assessment Detects power plant degradation for lower energy costs, higher reliability, and optimized maintenance Features Computes actual plant performance using ISO, DIN, IEC, and ISO performance test standards Compares actual performance to the performance expected under similar operating conditions Utilizes rigorous component models for performance prediction Models include boilers, turbines, pumps, fans, compressors, chillers, etc. Quantifies the observed degradation in terms of additional operating costs, or lost revenues Slide 26
Energy Efficiency and Performance Assessment Benefits Earlier detection of mechanical problems, preventing development of more serious damage Optimal maintenance scheduling Prevention of unscheduled shut-downs Accurate and up-to-date component models available for optimization Actual performance of the plant depicted, rather than the changes in operating conditions Slide 27
Energy Efficiency and Performance Assessment Cost Calculations Non optimized Operation is mirrored by Cost Deviations Focus to Problems according their relevance of Costs Higher Efficiency Better Cost Control Slide 28
Combustion Instruments Coal Flow - monitoring Advanced Control and Optimization of Power Plants Features On-line Pulverized Flow (PF) measurements of: Absolute PF velocity Burner PF split Relative PF loading (concentration) Mass flow rate, computed for each line, from split, but requires external total mass input (mill feed rate or similar) Benefits Measurement across total cross section Non-intrusive, low maintenance Virtually unaffected by roping Tolerant to poor upstream flow conditions No need for on-site calibration, unaffected by changes in coal type and moisture content Saves PF sampling cost Slide 29
Combustion Instruments Carbon In Ash - monitoring Features Online microwave measurement of unburnt coal in boiler backpass High accuracy Low maintenance Independent of coal quality and unit load Benefits Detection of arising problems with Burners Air adjustment / distribution Mills Improves ash quality Feedback or feed-forward signal for APC solutions Slide 30
Flame Scanners Advanced Control and Optimization of Power Plants Combustion Instruments Features Detects individual flame presence Provides flame quality index Continuous data to any plant control system via standard interfaces Individual flame raw signal for flame analysis and burner diagnosis Benefits Improve safety, reliability, reduce downtime Feedback or feed-forward signal for APC solutions Slide 31
D e t a i l e d Phase I Opportunity Assessment Three-Phase Delivery Process Phase II Implementation Phase III Follow-Up Objective: Identify opportunities and solutions Estimate benefits Justification Establish Baseline Proposal Objective: Engineering of the solution Audit Commissioning Training Objective: Verification of savings Sustaining the results Slide 32
Conclusions can deliver a full scope of Energy Management solutions to the Power Industry s Energy Management portfolio consists of different layers of instrumentation, control and optimization solutions Three-Phase Delivery Process System-approach Coordinated solutions between different sub-systems Multivariable Model Predictive Control (MPC) as underlying technology for advanced control Benefits: Better plant stability and availability Improved agility (faster start-up times and ramp rates, increased capacity, and reduced stable minimum loads) Higher plant efficiency Reduced emissions Slide 33
Slide 34